The purpose of this project was to design and build a new database, organised around Patients, to hold the information received by CDSC about positive HIV tests and AIDS diagnoses. This replaced two separate systems for HIV and AIDS which had been built up by the research staff of the section over the previous decade.
The system is implemented in Microsoft Access, on PCs running Windows 95, across a network running Windows NT. It currently holds information for about 30,000 Patients. New information is input from paper forms, and from batches supplied on file from larger reporting centres. The system provides for double entry of critical fields, and for matching new reports against the existing patients. The latter is done by computing a matching score over important fields, using this to order potential matches between existing patients and the new report, with the operator making the final decision to select a match or not. The operator is able to browse through all the information currently stored for a patient, and the matching decision is itself verified through repetition.
On a regular basis (usually every quarter) information about all patients is extracted to a separate database (one for each period) in a simplified and recoded form, for use in reporting and analysis. Aggregations are computed within MS Access, but the aggregates are passed to MS Excel for detailed layout and formatting as camera-ready output.
The system has a separate component for statistical meta-data, particularly labels and descriptions for fields and codes, treatment of special values, and recoding. It is used to control the codes available for input into fields, as a ‘Help’ system during entry, to provide labels for reports, and in the generation of derived variables. This system and the information it contains has the potential to be shared with other applications. Further details can be found in the Meta-Data section.
The researchers in the section have been trained in the use of MS Access, and have quickly become familiar with the construction of queries and reports to persue their own analyses on the database and on the extracts.
Our involvement with this project arose because the research section was having difficulties communicating its requirements for a database to the CDSC IT division, who in turn had problems in appreciating the requirements of a research database. After a brief initial consultation to identify the underlying problems and requirements we were invited to undertake the complete development.
Page last updated 29 April 1997.